In long term evolution (LTE) downlink multiple-input multiple-output (MIMO) transmission modes, each user equipment node (UE) feeds back the “best” rank indicator (RI), precoding matrix indicator (PMI), or both to a base station (BS), where “best” is determined according to the UE's criterion of maximizing its own throughput. However, “maximizing its own throughput” at every UE is not equivalent to “maximizing overall system throughput”, because different RI or PMI feedback can result in different interferences to other UEs in the network. This is particularly true when MIMO vertical tilting is used, because beams that are tilted down at a larger angle generally cause less interference to other UEs. Beam coordination between cells can be employed to help maximize overall system throughput. However, conventional systems do not consider beam coordination between cells. Dynamic “beam coordination” among nearby UEs requires complicated algorithms and real-time message exchanges among BSs, and increases overhead and system complexity.